Early days of technology employed traditional methods like rainfall distribution, duration ofrnsunshine, and the related meteorological data to forecast a forthcoming event regardingrndrought. Recently, satellite remote sensing has been considered as an appropriate tool forrnderiving information in spatial and temporal domains by providing multi-spectral reflectancerndata at regular intervals. Satellites from centers like National Oceanic and AtmosphericrnAdministration (NOAA) and Meteosat capture the spectral reflectance from green plant andrnprovide this frequently to monitor green vegetation conditions on the ground. Using the red andrninfrared band reflectances a vegetation index called Normalized Difference Vegetation Indexrn(NDVI) was derived; which is vital to access the evolution of drought as well as predict croprnyield.rnThe aims of this study are to analyze series of deviation of NDVI images, extract virtualrndrought objects from the series, and investigate for drought patterns from historical image forrngrowing season. Subsequent to this, appropriate prediction model of these patterns wasrndeveloped for early measures while within the same season. And it was applied on the subset ofrnimage data with reported drought occurrences in Ethiopia.rnIn this study, the virtual drought objects extracted from images over the growing season (June -rnSeptember) were found to exhibit a given pattern for the historical drought years. Afterrnproducing the descriptors of drought objects in the series using principal component analysis,rncombined and separate artificial neural network (ANN) models were used to predict thesernpatterns. In the combined approach all the descriptors of the object in the next time step werernpredicted all at the same time while in the separate approach the prediction was made one byrnone. Accordingly, the models designed to forecast the future state of drought object using theserntwo approaches yielded promising results. Especially, the three and four time lag combinedrnANN prediction model produced an overall RMSE of 20.80 and 16.43, respectively, which wasrna better result compared to a 36.92 RMSE of separate ANN approach. It is understood that thisrnwork will give new views for ways in drought prediction for early warning and crop conditionrnmonitoring at near real-time.rnKeywords: Drought prediction, NDVI images, Virtual drought object, Intelligent system